Determinants of successful cognitive aging over 15 years: a population‐based study

Abstract

AbstractBackgroundAlthough aging is related to loss in cognitive function, some individuals maintain high levels of cognitive performance also in late life. However, studies mapping longitudinal cognitive trajectories are scarce and little is known about the factors that may promote successful cognitive aging (SCA). We sought to identify predictors of SCA, based on latent profiles derived from interindividual differences in rate of change across several cognitive domains.MethodWe followed 1,966 individuals from the Swedish National Study on Aging and Care in Kungsholmen (SNAC‐K) over 15 years (mean baseline age = 71.6 years; 62.9% female). Participants were free from dementia at baseline and had undergone at least two cognitive assessments during follow‐up. Cognition was assessed with a comprehensive battery, including tests of semantic memory, episodic recall and recognition, letter and verbal fluency, and executive function. We estimated level and change within each of the cognitive domains with linear mixed effect models, based on which we classified participants into successful, normal, or decline groups. These groups were used to derive profiles of overall cognitive change trajectories, through latent class analysis. We then investigated demographic, health, lifestyle, and genetic predictors of profile membership.ResultWe obtained three latent cognitive profiles. 31.6% of the sample was classified as successful, 50.6% as normal, and 17.8% as decliner. In multi‐adjusted multinomial logistic regression analysis, membership in the successful profile was predicted by female sex, higher education, and a faster walking speed (ps < 0.003). Current smoking, feelings of loneliness, and carrying the Apolipoprotein ԑ4‐allele were associated with a lower likelihood of being classified as successful (ps < 0.037). Higher age, diagnosis of diabetes or depression, and being an ԑ4‐carrier increased the likelihood of being classified into the decline profile (ps < 0.026).ConclusionWe identified several demographic, lifestyle, health, and genetic factors of SCA. Of note, these factors were different from those that predicted accelerated cognitive decline. Although some cognitive loss may be an inevitable part of aging, our results suggest certain modifiable factors (i.e. smoking and quality of social interactions) to be related to maintained cognitive function in older age.